PURPOSE: To determine the accuracy of frequency compounding (FC) in detection and characterisation of breast lesions by ultrasound (US) and to assess the image quality in direct comparison with tissue harmonic imaging (THI). MATERIALS AND METHODS: A total of 60 breast lesions were examined; these were malignant in 39 cases and benign in 21 as confirmed by histology of core biopsy specimens (Aplio 80, Toshiba, Otawara, Japan). A total of 120 freeze frames, 60 B-mode scans with THI and 60 scans with FC were assessed by three independent examiners, who determined the presence of pathology on a scale of 1 to 5 using established criteria (BI-RADS classification). Image quality of each technique was rated on a ranking scale of - 2 (FC clearly superior to THI), - 1 (FC slightly superior to THI), 0 (identical), + 1 (THI slightly superior to FC), to + 2 (THI clearly superior to FC). The different US techniques were compared in terms of image quality and diagnostic accuracy using Mann Whitney U test and ROC analysis. RESULTS: The results for image quality were as follows: - 0.56 +/- 0.71. In 56 % of the images FC was superior to THI, both techniques were identical in 37 % of cases, and in only 7 % THI was superior to FC. The following AUCs (Area under Curve, presence of pathology) were calculated: 0.945 and 0.969 for THI and FC, respectively (not significant). The two US techniques had comparable sensitivity (FC/THI 100/98 %) and specificity (FC/THI 76/74 %). CONCLUSION: Altogether, the diagnostic value of FC was comparable to that of THI with only slight differences. For ultrasound examinations of breast lesions, FC improves the image quality and conspicuity of pathology.
PURPOSE: To determine the accuracy of frequency compounding (FC) in detection and characterisation of breast lesions by ultrasound (US) and to assess the image quality in direct comparison with tissue harmonic imaging (THI). MATERIALS AND METHODS: A total of 60 breast lesions were examined; these were malignant in 39 cases and benign in 21 as confirmed by histology of core biopsy specimens (Aplio 80, Toshiba, Otawara, Japan). A total of 120 freeze frames, 60 B-mode scans with THI and 60 scans with FC were assessed by three independent examiners, who determined the presence of pathology on a scale of 1 to 5 using established criteria (BI-RADS classification). Image quality of each technique was rated on a ranking scale of - 2 (FC clearly superior to THI), - 1 (FC slightly superior to THI), 0 (identical), + 1 (THI slightly superior to FC), to + 2 (THI clearly superior to FC). The different US techniques were compared in terms of image quality and diagnostic accuracy using Mann Whitney U test and ROC analysis. RESULTS: The results for image quality were as follows: - 0.56 +/- 0.71. In 56 % of the images FC was superior to THI, both techniques were identical in 37 % of cases, and in only 7 % THI was superior to FC. The following AUCs (Area under Curve, presence of pathology) were calculated: 0.945 and 0.969 for THI and FC, respectively (not significant). The two US techniques had comparable sensitivity (FC/THI 100/98 %) and specificity (FC/THI 76/74 %). CONCLUSION: Altogether, the diagnostic value of FC was comparable to that of THI with only slight differences. For ultrasound examinations of breast lesions, FC improves the image quality and conspicuity of pathology.